Presentation is loading. Please wait.

Presentation is loading. Please wait.

Dünaamiliste süsteemide modelleerimine Identification for control in a non- linear system world Eduard Petlenkov, Automaatikainstituut, 2013.

Similar presentations


Presentation on theme: "Dünaamiliste süsteemide modelleerimine Identification for control in a non- linear system world Eduard Petlenkov, Automaatikainstituut, 2013."— Presentation transcript:

1 Dünaamiliste süsteemide modelleerimine Identification for control in a non- linear system world Eduard Petlenkov, Automaatikainstituut, 2013

2 Control Design Nominal Model Regulator © Lennart Ljung

3 Control Design True System Regulator © Lennart Ljung

4 Control Design Nominal Model Regulator Model error model © Lennart Ljung

5 Control Design Nominal Model Regulator Model error model True system © Lennart Ljung

6 Control Design Nominal Model Regulator Model error model Nominal closed loop system © Lennart Ljung

7 Model Error Models   = y – y model u  © Lennart Ljung

8 Identification for Control Identifiction for control is the art and technique to design identification experiments and regulator design methods so that the model error model matches the nominal closed loop system in a suitable way © Lennart Ljung

9 A Data Set/ nonlinear process Output Input © Lennart Ljung

10 Feedback Neural Networks. Identification of dynamic systems (1) Eduard Petlenkov, Automaatikainstituut, 2013 External feedback:

11 11 Eduard Petlenkov, Automaatikainstituut, 2013 Elman’i võrk Internal feedback: where r is the number of neurons on the recurrent layer; Feedback Neural Networks. Identification of dynamic systems (2)

12 Identification with ANNs (1) Eduard Petlenkov, Automaatikainstituut, 2013 is the input of the system is the output of the system is the output of the model

13 Identification with ANNs (2) Eduard Petlenkov, Automaatikainstituut, 2013

14 Identification with ANNs (3) Eduard Petlenkov, Automaatikainstituut, 2013 1. Getting experimental training data: Input signal is given to the system and the corresponding output reaction is measured. Obtained input-output data is used to train a NN-based model; 2. Choosing NN’s structure: number of inputs, outputs, hidden layers, neurons; activation functions; 3. Choosing initial weighting coefficients (usually random).

15 Identification with ANNs (3) Eduard Petlenkov, Automaatikainstituut, 2013 4. Calculation of NN outputs corresponding to the training input set; 5. Comparison of outputs with reference outputs from the training set – calculation of model error; 6. Calculation of uptadeted weighting coefficients using selected training algorithm; Repeat steps 4-6 untill a predefined number of iterations (epochs) is reached or until the desire accuracy is achieved.

16 Dünaamiliste süsteemide juhtimine tehisnärvivõrkudega NN-based control of dynamic systems Eduard Petlenkov, Automaatikainstituut, 2013

17 Predictive Control

18 Inverse model Eduard Petlenkov, Automaatikainstituut, 2013

19 Inverse model Eduard Petlenkov, Automaatikainstituut, 2013

20 Inverse model based control Here Eduard Petlenkov, Automaatikainstituut, 2013

21 Adaptive inverse model based control Eduard Petlenkov, Automaatikainstituut, 2013

22 Adaptive Predictive Control

23 Mustrite tuvastamine ja klassifitseerimine tehisnärvivõrkudega NN-based pattern recognition and classification Eduard Petlenkov, Automaatikainstituut, 2013

24 24 Self-organizing networks Self-orgamizing network is capable of adjusting its parameters on the basis of inputs and chosen optimization criteria. It doesn’t have any reference outputs Kohonen map Eduard Petlenkov, Automaatikainstituut, 2013

25 Näide 1: Character recognition(1)

26 Näide 1: Character recognition (2)

27 Näide 2: number recognition

28 Etalon_1=[1 0 0 0 0 0 0 0 0 0]'; Etalon_2=[0 1 0 0 0 0 0 0 0 0]'; Etalon_3=[0 0 1 0 0 0 0 0 0 0]'; Etalon_4=[0 0 0 1 0 0 0 0 0 0]'; Etalon_5=[0 0 0 0 1 0 0 0 0 0]'; Etalon_6=[0 0 0 0 0 1 0 0 0 0]'; Etalon_7=[0 0 0 0 0 0 1 0 0 0]'; Etalon_8=[0 0 0 0 0 0 0 1 0 0]'; Etalon_9=[0 0 0 0 0 0 0 0 1 0]'; Etalon_0=[0 0 0 0 0 0 0 0 0 1]'; net = newff(minmax(P),[25 10],{'logsig' 'logsig'},'traingda') net.trainParam.goal = 0 net.trainParam.epochs=10000 Näide 2: Number recognition– NN structure

29 Näide 2: Number recognition– testing on the training set 1234567890 10.99990.0000 2 0.99990.0000 3 0.99990.0000 4 0.99990.0000 5 0.99990.0000 6 0.99990.0000 7 0.99990.0000 8 0.99990.0000 9 0.99990.0000 100.0000 0.9999

30 Näide 2: Number recognition– testing on the validation set 1234567890 1 0,06680,00060,00660,00260,00890,01520,00340,00000,00500,0001 2 0,00190,15480,00140,00050,01080,00050,00440,00010,00000,0013 3 0,01450,00020,14950,00010,00320,03430,00810,00010,00390,0001 4 0,00030,00260,00010,01020,00040,00000,00140,00470,00080,0092 5 0,00000,00030,0000 0,01770,0000 6 0,00010,00130,00200,00020,00050,06110,00020,00000,00650,0081 7 0,00020,00000,00020,00030,00010,00000,01540,00090,0000 8 0,10300,0000 0,58680,0000 9 0,0002 0,00030,00080,05800,00560,00060,00100,27990,0051 10 0,0000 0,0186

31 Databases for benchmark problems  MNIST database of handwritten digits  26x26px  60 000 labeled images and 10 000 test images  Semeion - 1593 handwritten digits from around 80 persons  16x16px


Download ppt "Dünaamiliste süsteemide modelleerimine Identification for control in a non- linear system world Eduard Petlenkov, Automaatikainstituut, 2013."

Similar presentations


Ads by Google